Merge branch 'mirabilos-asm-elf-pie' into devel

9 files changed, 414 insertions, 353 deletions

diff --git a/src/amd64/Makefile.am b/src/amd64/Makefile.am
index 07aae69..2870ab5 100644
--- a/src/amd64/Makefile.am
+++ b/src/amd64/Makefile.am
@@ -1,3 +1,5 @@
+NAFLAGS += -DASM_ARCH_AMD64
+
 AMD64_ASM = \
   cpuid_amd64.asm \
   rfxcodec_encode_dwt_shift_amd64_sse2.asm \
diff --git a/src/amd64/cpuid_amd64.asm b/src/amd64/cpuid_amd64.asm
index acc738e..38e2023 100644
--- a/src/amd64/cpuid_amd64.asm
+++ b/src/amd64/cpuid_amd64.asm
@@ -1,7 +1,5 @@
 %include "common.asm"
 
-section .text
-
 ;The first six integer or pointer arguments are passed in registers
 ;RDI, RSI, RDX, RCX, R8, and R9
 
@@ -32,5 +30,4 @@ PROC cpuid_amd64
     ; restore registers
     pop rbx
     ret
-    align 16
-
+END_OF_FILE
diff --git a/src/amd64/rfxcodec_encode_dwt_shift_amd64_sse2.asm b/src/amd64/rfxcodec_encode_dwt_shift_amd64_sse2.asm
index 69ccf07..cef3902 100644
--- a/src/amd64/rfxcodec_encode_dwt_shift_amd64_sse2.asm
+++ b/src/amd64/rfxcodec_encode_dwt_shift_amd64_sse2.asm
@@ -21,8 +21,7 @@
 
 %include "common.asm"
 
-section .data
-    align 16
+PREPARE_RODATA
     cw128    times 8 dw 128
     cdFFFF   times 4 dd 65535
     ; these are 1 << (factor - 1) 0 to 15 is factor
@@ -43,8 +42,6 @@ section .data
     cwa8192  times 8 dw 8192  ; 14
     cwa16384 times 8 dw 16384 ; 15
 
-section .text
-
 ;******************************************************************************
 ; source 16 bit signed, 16 pixel width
 rfx_dwt_2d_encode_block_horiz_16_16:
@@ -55,8 +52,8 @@ loop1a:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -66,8 +63,8 @@ loop1a:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -85,8 +82,8 @@ loop1a:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -247,8 +244,8 @@ loop1c:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -258,8 +255,8 @@ loop1c:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -277,8 +274,8 @@ loop1c:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -328,8 +325,8 @@ loop1c:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -339,8 +336,8 @@ loop1c:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -358,8 +355,8 @@ loop1c:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -423,8 +420,8 @@ loop1c1:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -434,8 +431,8 @@ loop1c1:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -453,8 +450,8 @@ loop1c1:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -501,8 +498,8 @@ loop1c1:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -512,8 +509,8 @@ loop1c1:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -531,8 +528,8 @@ loop1c1:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -690,8 +687,8 @@ loop1e:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -701,8 +698,8 @@ loop1e:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -720,8 +717,8 @@ loop1e:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -774,8 +771,8 @@ loop2e:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -785,8 +782,8 @@ loop2e:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -804,8 +801,8 @@ loop2e:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -857,8 +854,8 @@ loop2e:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -868,8 +865,8 @@ loop2e:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -887,8 +884,8 @@ loop2e:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -952,8 +949,8 @@ loop1e1:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -963,8 +960,8 @@ loop1e1:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -982,8 +979,8 @@ loop1e1:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -1033,8 +1030,8 @@ loop2e1:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -1044,8 +1041,8 @@ loop2e1:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -1063,8 +1060,8 @@ loop2e1:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -1113,8 +1110,8 @@ loop2e1:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -1124,8 +1121,8 @@ loop2e1:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -1143,8 +1140,8 @@ loop2e1:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -1207,9 +1204,9 @@ loop1f:
     punpcklbw xmm1, xmm0
     punpcklbw xmm2, xmm0
     punpcklbw xmm3, xmm0
-    psubw xmm1, [rel cw128]
-    psubw xmm2, [rel cw128]
-    psubw xmm3, [rel cw128]
+    psubw xmm1, [lsym(cw128)]
+    psubw xmm2, [lsym(cw128)]
+    psubw xmm3, [lsym(cw128)]
     psllw xmm1, 5
     psllw xmm2, 5
     psllw xmm3, 5
@@ -1241,8 +1238,8 @@ loop2f:
     movq xmm3, [rsi + 64 * 1 * 2]       ; src[2n + 2]
     punpcklbw xmm2, xmm0
     punpcklbw xmm3, xmm0
-    psubw xmm2, [rel cw128]
-    psubw xmm3, [rel cw128]
+    psubw xmm2, [lsym(cw128)]
+    psubw xmm3, [lsym(cw128)]
     psllw xmm2, 5
     psllw xmm3, 5
     movdqa xmm4, xmm1
@@ -1274,7 +1271,7 @@ loop2f:
     movdqa xmm1, xmm3                   ; src[2n]
     movq xmm2, [rsi + 64 * 1]           ; src[2n + 1]
     punpcklbw xmm2, xmm0
-    psubw xmm2, [rel cw128]
+    psubw xmm2, [lsym(cw128)]
     psllw xmm2, 5
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -1314,7 +1311,7 @@ set_quants_hi:
     sub rax, 6 - 5
     movd xmm9, eax
     imul rax, 16
-    lea rdx, [rel cwa0]
+    lea rdx, [lsym(cwa0)]
     add rdx, rax
     movdqa xmm8, [rdx]
     ret
@@ -1323,7 +1320,7 @@ set_quants_lo:
     sub rax, 6 - 5
     movd xmm11, eax
     imul rax, 16
-    lea rdx, [rel cwa0]
+    lea rdx, [lsym(cwa0)]
     add rdx, rax
     movdqa xmm10, [rdx]
     ret
@@ -1487,5 +1484,4 @@ PROC rfxcodec_encode_dwt_shift_amd64_sse2
     pop rdx
     pop rbx
     ret
-    align 16
-
+END_OF_FILE
diff --git a/src/amd64/rfxcodec_encode_dwt_shift_amd64_sse41.asm b/src/amd64/rfxcodec_encode_dwt_shift_amd64_sse41.asm
index 2b19f81..da176e7 100644
--- a/src/amd64/rfxcodec_encode_dwt_shift_amd64_sse41.asm
+++ b/src/amd64/rfxcodec_encode_dwt_shift_amd64_sse41.asm
@@ -21,8 +21,7 @@
 
 %include "common.asm"
 
-section .data
-    align 16
+PREPARE_RODATA
     cw128    times 8 dw 128
     cdFFFF   times 4 dd 65535
     ; these are 1 << (factor - 1) 0 to 15 is factor
@@ -43,8 +42,6 @@ section .data
     cwa8192  times 8 dw 8192  ; 14
     cwa16384 times 8 dw 16384 ; 15
 
-section .text
-
 ;******************************************************************************
 ; source 16 bit signed, 16 pixel width
 rfx_dwt_2d_encode_block_horiz_16_16:
@@ -55,15 +52,15 @@ loop1a:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -77,8 +74,8 @@ loop1a:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -235,15 +232,15 @@ loop1c:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -257,8 +254,8 @@ loop1c:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -304,15 +301,15 @@ loop1c:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -326,8 +323,8 @@ loop1c:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -387,15 +384,15 @@ loop1c1:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -409,8 +406,8 @@ loop1c1:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -453,15 +450,15 @@ loop1c1:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -475,8 +472,8 @@ loop1c1:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -630,15 +627,15 @@ loop1e:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -652,8 +649,8 @@ loop1e:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -702,15 +699,15 @@ loop2e:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -724,8 +721,8 @@ loop2e:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -773,15 +770,15 @@ loop2e:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -795,8 +792,8 @@ loop2e:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -856,15 +853,15 @@ loop1e1:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -878,8 +875,8 @@ loop1e1:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -925,15 +922,15 @@ loop2e1:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -947,8 +944,8 @@ loop2e1:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -993,15 +990,15 @@ loop2e1:
     movdqa xmm2, [rsi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [rel cdFFFF]
-    pand xmm2, [rel cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [rel cdFFFF]
-    pand xmm3, [rel cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -1015,8 +1012,8 @@ loop2e1:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [rel cdFFFF]
-    pand xmm4, [rel cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -1075,9 +1072,9 @@ loop1f:
     punpcklbw xmm1, xmm0
     punpcklbw xmm2, xmm0
     punpcklbw xmm3, xmm0
-    psubw xmm1, [rel cw128]
-    psubw xmm2, [rel cw128]
-    psubw xmm3, [rel cw128]
+    psubw xmm1, [lsym(cw128)]
+    psubw xmm2, [lsym(cw128)]
+    psubw xmm3, [lsym(cw128)]
     psllw xmm1, 5
     psllw xmm2, 5
     psllw xmm3, 5
@@ -1109,8 +1106,8 @@ loop2f:
     movq xmm3, [rsi + 64 * 1 * 2]       ; src[2n + 2]
     punpcklbw xmm2, xmm0
     punpcklbw xmm3, xmm0
-    psubw xmm2, [rel cw128]
-    psubw xmm3, [rel cw128]
+    psubw xmm2, [lsym(cw128)]
+    psubw xmm3, [lsym(cw128)]
     psllw xmm2, 5
     psllw xmm3, 5
     movdqa xmm4, xmm1
@@ -1142,7 +1139,7 @@ loop2f:
     movdqa xmm1, xmm3                   ; src[2n]
     movq xmm2, [rsi + 64 * 1]           ; src[2n + 1]
     punpcklbw xmm2, xmm0
-    psubw xmm2, [rel cw128]
+    psubw xmm2, [lsym(cw128)]
     psllw xmm2, 5
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -1182,7 +1179,7 @@ set_quants_hi:
     sub rax, 6 - 5
     movd xmm9, eax
     imul rax, 16
-    lea rdx, [rel cwa0]
+    lea rdx, [lsym(cwa0)]
     add rdx, rax
     movdqa xmm8, [rdx]
     ret
@@ -1191,7 +1188,7 @@ set_quants_lo:
     sub rax, 6 - 5
     movd xmm11, eax
     imul rax, 16
-    lea rdx, [rel cwa0]
+    lea rdx, [lsym(cwa0)]
     add rdx, rax
     movdqa xmm10, [rdx]
     ret
@@ -1355,5 +1352,4 @@ PROC rfxcodec_encode_dwt_shift_amd64_sse41
     pop rdx
     pop rbx
     ret
-    align 16
-
+END_OF_FILE
diff --git a/src/common.asm b/src/common.asm
index 7e2b84b..cf7102d 100644
--- a/src/common.asm
+++ b/src/common.asm
@@ -1,5 +1,6 @@
 ;
 ;Copyright 2017 Pavel Roskin
+;Copyright 2017 mirabilos
 ;
 ;Permission to use, copy, modify, distribute, and sell this software and its
 ;documentation for any purpose is hereby granted without fee, provided that
@@ -33,6 +34,19 @@
 %define is_elf 1
 %endif
 
+; Detect Mach-O formats
+%ifidn __OUTPUT_FORMAT__,macho
+%define is_macho 1
+%endif
+
+%ifidn __OUTPUT_FORMAT__,macho32
+%define is_macho 1
+%endif
+
+%ifidn __OUTPUT_FORMAT__,macho64
+%define is_macho 1
+%endif
+
 ; Mark stack non-executable
 %ifdef is_elf
 section .note.GNU-stack noalloc noexec nowrite progbits
@@ -49,3 +63,64 @@ section .note.GNU-stack noalloc noexec nowrite progbits
     _%1:
 %endif
 %endmacro
+
+; Macros for relative access to local data
+%undef lsym
+
+%ifdef ASM_ARCH_AMD64
+; amd64; don't define or call RETRIEVE_RODATA
+%define lsym(name) rel name
+%endif
+
+%ifdef ASM_ARCH_I386
+%ifdef PIC
+; i386 PIC
+
+%macro END_OF_FILE 0
+%ifdef I386_PIC_NEEDED
+section .text
+..@get_caller_address:
+	mov ebx, [esp]
+	ret
+%endif
+%ifdef is_macho
+; see below
+    align 16
+%endif
+%endmacro
+
+%macro RETRIEVE_RODATA 0
+%define I386_PIC_NEEDED 1
+	call ..@get_caller_address
+%%the_caller_address:
+	sub ebx, %%the_caller_address - ..@rodata_begin
+%endmacro
+
+%define lsym(name) ebx + name - ..@rodata_begin
+%else
+; i386 non-PIC; default case for lsym and RETRIEVE_RODATA
+%endif
+%endif
+
+%ifndef lsym
+%macro RETRIEVE_RODATA 0
+%endmacro
+%define lsym(name) name
+%endif
+
+%macro PREPARE_RODATA 0
+section .text
+    align 16
+..@rodata_begin:
+%endmacro
+
+%ifnmacro END_OF_FILE 0
+%macro END_OF_FILE 0
+%ifdef is_macho
+; cf. https://github.com/libjpeg-turbo/libjpeg-turbo/blob/master/simd/jccolext-mmx.asm#L474-L476
+    align 16
+%endif
+%endmacro
+%endif
+
+section .text
diff --git a/src/x86/Makefile.am b/src/x86/Makefile.am
index 2d099e5..3c88cee 100644
--- a/src/x86/Makefile.am
+++ b/src/x86/Makefile.am
@@ -1,3 +1,5 @@
+NAFLAGS += -DASM_ARCH_I386
+
 X86_ASM = \
   cpuid_x86.asm \
   rfxcodec_encode_dwt_shift_x86_sse2.asm \
diff --git a/src/x86/cpuid_x86.asm b/src/x86/cpuid_x86.asm
index 4ddb8a2..b666732 100644
--- a/src/x86/cpuid_x86.asm
+++ b/src/x86/cpuid_x86.asm
@@ -1,7 +1,5 @@
 %include "common.asm"
 
-section .text
-
 ;int
 ;cpuid_x86(int eax_in, int ecx_in, int *eax, int *ebx, int *ecx, int *edx)
 
@@ -29,6 +27,5 @@ PROC cpuid_x86
     pop edx
     pop ecx
     pop ebx
-    ret;
-    align 16
-
+    ret
+END_OF_FILE
diff --git a/src/x86/rfxcodec_encode_dwt_shift_x86_sse2.asm b/src/x86/rfxcodec_encode_dwt_shift_x86_sse2.asm
index fdfbae1..f05a705 100644
--- a/src/x86/rfxcodec_encode_dwt_shift_x86_sse2.asm
+++ b/src/x86/rfxcodec_encode_dwt_shift_x86_sse2.asm
@@ -1,5 +1,6 @@
 ;
 ;Copyright 2016 Jay Sorg
+;Copyright 2017 mirabilos
 ;
 ;Permission to use, copy, modify, distribute, and sell this software and its
 ;documentation for any purpose is hereby granted without fee, provided that
@@ -21,8 +22,7 @@
 
 %include "common.asm"
 
-section .data
-    align 16
+PREPARE_RODATA
     cw128    times 8 dw 128
     cdFFFF   times 4 dd 65535
     ; these are 1 << (factor - 1) 0 to 15 is factor
@@ -43,8 +43,6 @@ section .data
     cwa8192  times 8 dw 8192  ; 14
     cwa16384 times 8 dw 16384 ; 15
 
-section .text
-
 %define LHI_ADD  [esp + 1 * 16 + 4]
 %define LHI_SFT  [esp + 2 * 16 + 4]
 %define LLO_ADD  [esp + 3 * 16 + 4]
@@ -60,8 +58,8 @@ loop1a:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -71,8 +69,8 @@ loop1a:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -90,8 +88,8 @@ loop1a:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -252,8 +250,8 @@ loop1c:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -263,8 +261,8 @@ loop1c:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -282,8 +280,8 @@ loop1c:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -316,7 +314,7 @@ loop1c:
     paddw xmm5, xmm1
 
     psrldq xmm2, 14
-    movd ebx, xmm2                      ; save hi
+    movd ebp, xmm2                      ; save hi
 
     movdqa xmm6, xmm5                   ; out lo
     paddw xmm6, LLO_ADD
@@ -333,8 +331,8 @@ loop1c:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -344,8 +342,8 @@ loop1c:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -363,8 +361,8 @@ loop1c:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -387,7 +385,7 @@ loop1c:
     ; l[n] = src[2n] + ((h[n - 1] + h[n]) >> 1)
     movdqa xmm7, xmm5
     pslldq xmm7, 2
-    movd xmm6, ebx
+    movd xmm6, ebp
     por xmm7, xmm6
     paddw xmm5, xmm7
     psraw xmm5, 1
@@ -428,8 +426,8 @@ loop1c1:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -439,8 +437,8 @@ loop1c1:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -458,8 +456,8 @@ loop1c1:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -492,7 +490,7 @@ loop1c1:
     paddw xmm5, xmm1
 
     psrldq xmm2, 14
-    movd ebx, xmm2                      ; save hi
+    movd ebp, xmm2                      ; save hi
 
     movdqa [edx], xmm5                  ; out lo
 
@@ -506,8 +504,8 @@ loop1c1:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -517,8 +515,8 @@ loop1c1:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -536,8 +534,8 @@ loop1c1:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -560,7 +558,7 @@ loop1c1:
     ; l[n] = src[2n] + ((h[n - 1] + h[n]) >> 1)
     movdqa xmm7, xmm5
     pslldq xmm7, 2
-    movd xmm6, ebx
+    movd xmm6, ebp
     por xmm7, xmm6
     paddw xmm5, xmm7
     psraw xmm5, 1
@@ -695,8 +693,8 @@ loop1e:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -706,8 +704,8 @@ loop1e:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -725,8 +723,8 @@ loop1e:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -759,7 +757,7 @@ loop1e:
     paddw xmm5, xmm1
 
     psrldq xmm2, 14
-    movd ebx, xmm2                      ; save hi
+    movd ebp, xmm2                      ; save hi
 
     movdqa xmm6, xmm5                   ; out lo
     paddw xmm6, LLO_ADD
@@ -779,8 +777,8 @@ loop2e:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -790,8 +788,8 @@ loop2e:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -809,8 +807,8 @@ loop2e:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -834,14 +832,14 @@ loop2e:
     ; l[n] = src[2n] + ((h[n - 1] + h[n]) >> 1)
     movdqa xmm7, xmm5
     pslldq xmm7, 2
-    movd xmm6, ebx
+    movd xmm6, ebp
     por xmm7, xmm6
     paddw xmm5, xmm7
     psraw xmm5, 1
     paddw xmm5, xmm1
 
     psrldq xmm2, 14
-    movd ebx, xmm2                      ; save hi
+    movd ebp, xmm2                      ; save hi
 
     movdqa xmm6, xmm5                   ; out lo
     paddw xmm6, LLO_ADD
@@ -862,8 +860,8 @@ loop2e:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -873,8 +871,8 @@ loop2e:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -892,8 +890,8 @@ loop2e:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -916,7 +914,7 @@ loop2e:
     ; l[n] = src[2n] + ((h[n - 1] + h[n]) >> 1)
     movdqa xmm7, xmm5
     pslldq xmm7, 2
-    movd xmm6, ebx
+    movd xmm6, ebp
     por xmm7, xmm6
     paddw xmm5, xmm7
     psraw xmm5, 1
@@ -957,8 +955,8 @@ loop1e1:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -968,8 +966,8 @@ loop1e1:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -987,8 +985,8 @@ loop1e1:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -1021,7 +1019,7 @@ loop1e1:
     paddw xmm5, xmm1
 
     psrldq xmm2, 14
-    movd ebx, xmm2                      ; save hi
+    movd ebp, xmm2                      ; save hi
 
     movdqa [edx], xmm5                  ; out lo
 
@@ -1038,8 +1036,8 @@ loop2e1:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -1049,8 +1047,8 @@ loop2e1:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -1068,8 +1066,8 @@ loop2e1:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -1093,14 +1091,14 @@ loop2e1:
     ; l[n] = src[2n] + ((h[n - 1] + h[n]) >> 1)
     movdqa xmm7, xmm5
     pslldq xmm7, 2
-    movd xmm6, ebx
+    movd xmm6, ebp
     por xmm7, xmm6
     paddw xmm5, xmm7
     psraw xmm5, 1
     paddw xmm5, xmm1
 
     psrldq xmm2, 14
-    movd ebx, xmm2                      ; save hi
+    movd ebp, xmm2                      ; save hi
 
     movdqa [edx], xmm5                  ; out lo
 
@@ -1118,8 +1116,8 @@ loop2e1:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     pslld xmm1, 16
     pslld xmm2, 16
     psrad xmm1, 16
@@ -1129,8 +1127,8 @@ loop2e1:
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     pslld xmm2, 16
     pslld xmm3, 16
     psrad xmm2, 16
@@ -1148,8 +1146,8 @@ loop2e1:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     pslld xmm3, 16
     pslld xmm4, 16
     psrad xmm3, 16
@@ -1172,7 +1170,7 @@ loop2e1:
     ; l[n] = src[2n] + ((h[n - 1] + h[n]) >> 1)
     movdqa xmm7, xmm5
     pslldq xmm7, 2
-    movd xmm6, ebx
+    movd xmm6, ebp
     por xmm7, xmm6
     paddw xmm5, xmm7
     psraw xmm5, 1
@@ -1212,9 +1210,9 @@ loop1f:
     punpcklbw xmm1, xmm0
     punpcklbw xmm2, xmm0
     punpcklbw xmm3, xmm0
-    psubw xmm1, [cw128]
-    psubw xmm2, [cw128]
-    psubw xmm3, [cw128]
+    psubw xmm1, [lsym(cw128)]
+    psubw xmm2, [lsym(cw128)]
+    psubw xmm3, [lsym(cw128)]
     psllw xmm1, 5
     psllw xmm2, 5
     psllw xmm3, 5
@@ -1246,8 +1244,8 @@ loop2f:
     movq xmm3, [esi + 64 * 1 * 2]       ; src[2n + 2]
     punpcklbw xmm2, xmm0
     punpcklbw xmm3, xmm0
-    psubw xmm2, [cw128]
-    psubw xmm3, [cw128]
+    psubw xmm2, [lsym(cw128)]
+    psubw xmm3, [lsym(cw128)]
     psllw xmm2, 5
     psllw xmm3, 5
     movdqa xmm4, xmm1
@@ -1279,7 +1277,7 @@ loop2f:
     movdqa xmm1, xmm3                   ; src[2n]
     movq xmm2, [esi + 64 * 1]           ; src[2n + 1]
     punpcklbw xmm2, xmm0
-    psubw xmm2, [cw128]
+    psubw xmm2, [lsym(cw128)]
     psllw xmm2, 5
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -1320,7 +1318,7 @@ set_quants_hi:
     movd xmm1, eax
     movdqa LHI_SFT, xmm1
     imul eax, 16
-    lea edx, [cwa0]
+    lea edx, [lsym(cwa0)]
     add edx, eax
     movdqa xmm1, [edx]
     movdqa LHI_ADD, xmm1
@@ -1331,7 +1329,7 @@ set_quants_lo:
     movd xmm1, eax
     movdqa LLO_SFT, xmm1
     imul eax, 16
-    lea edx, [cwa0]
+    lea edx, [lsym(cwa0)]
     add edx, eax
     movdqa xmm1, [edx]
     movdqa LLO_ADD, xmm1
@@ -1363,6 +1361,7 @@ PROC rfxcodec_encode_dwt_shift_x86_sse2
     movdqu [esp], xmm0
     ; save registers
     push ebx
+    RETRIEVE_RODATA
     push esi
     push edi
     push ebp
@@ -1517,5 +1516,4 @@ PROC rfxcodec_encode_dwt_shift_x86_sse2
     ; return value
     mov eax, 0
     ret
-    align 16
-
+END_OF_FILE
diff --git a/src/x86/rfxcodec_encode_dwt_shift_x86_sse41.asm b/src/x86/rfxcodec_encode_dwt_shift_x86_sse41.asm
index 501a0bc..00d4b1d 100644
--- a/src/x86/rfxcodec_encode_dwt_shift_x86_sse41.asm
+++ b/src/x86/rfxcodec_encode_dwt_shift_x86_sse41.asm
@@ -1,5 +1,6 @@
 ;
 ;Copyright 2016 Jay Sorg
+;Copyright 2017 mirabilos
 ;
 ;Permission to use, copy, modify, distribute, and sell this software and its
 ;documentation for any purpose is hereby granted without fee, provided that
@@ -21,8 +22,7 @@
 
 %include "common.asm"
 
-section .data
-    align 16
+PREPARE_RODATA
     cw128    times 8 dw 128
     cdFFFF   times 4 dd 65535
     ; these are 1 << (factor - 1) 0 to 15 is factor
@@ -43,8 +43,6 @@ section .data
     cwa8192  times 8 dw 8192  ; 14
     cwa16384 times 8 dw 16384 ; 15
 
-section .text
-
 %define LHI_ADD  [esp + 1 * 16 + 4]
 %define LHI_SFT  [esp + 2 * 16 + 4]
 %define LLO_ADD  [esp + 3 * 16 + 4]
@@ -60,15 +58,15 @@ loop1a:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -82,8 +80,8 @@ loop1a:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -240,15 +238,15 @@ loop1c:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -262,8 +260,8 @@ loop1c:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -292,7 +290,7 @@ loop1c:
     paddw xmm5, xmm1
 
     psrldq xmm2, 14
-    movd ebx, xmm2                      ; save hi
+    movd ebp, xmm2                      ; save hi
 
     movdqa xmm6, xmm5                   ; out lo
     paddw xmm6, LLO_ADD
@@ -309,15 +307,15 @@ loop1c:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -331,8 +329,8 @@ loop1c:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -351,7 +349,7 @@ loop1c:
     ; l[n] = src[2n] + ((h[n - 1] + h[n]) >> 1)
     movdqa xmm7, xmm5
     pslldq xmm7, 2
-    movd xmm6, ebx
+    movd xmm6, ebp
     por xmm7, xmm6
     paddw xmm5, xmm7
     psraw xmm5, 1
@@ -392,15 +390,15 @@ loop1c1:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -414,8 +412,8 @@ loop1c1:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -444,7 +442,7 @@ loop1c1:
     paddw xmm5, xmm1
 
     psrldq xmm2, 14
-    movd ebx, xmm2                      ; save hi
+    movd ebp, xmm2                      ; save hi
 
     movdqa [edx], xmm5                  ; out lo
 
@@ -458,15 +456,15 @@ loop1c1:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -480,8 +478,8 @@ loop1c1:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -500,7 +498,7 @@ loop1c1:
     ; l[n] = src[2n] + ((h[n - 1] + h[n]) >> 1)
     movdqa xmm7, xmm5
     pslldq xmm7, 2
-    movd xmm6, ebx
+    movd xmm6, ebp
     por xmm7, xmm6
     paddw xmm5, xmm7
     psraw xmm5, 1
@@ -635,15 +633,15 @@ loop1e:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -657,8 +655,8 @@ loop1e:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -687,7 +685,7 @@ loop1e:
     paddw xmm5, xmm1
 
     psrldq xmm2, 14
-    movd ebx, xmm2                      ; save hi
+    movd ebp, xmm2                      ; save hi
 
     movdqa xmm6, xmm5                   ; out lo
     paddw xmm6, LLO_ADD
@@ -707,15 +705,15 @@ loop2e:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -729,8 +727,8 @@ loop2e:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -750,14 +748,14 @@ loop2e:
     ; l[n] = src[2n] + ((h[n - 1] + h[n]) >> 1)
     movdqa xmm7, xmm5
     pslldq xmm7, 2
-    movd xmm6, ebx
+    movd xmm6, ebp
     por xmm7, xmm6
     paddw xmm5, xmm7
     psraw xmm5, 1
     paddw xmm5, xmm1
 
     psrldq xmm2, 14
-    movd ebx, xmm2                      ; save hi
+    movd ebp, xmm2                      ; save hi
 
     movdqa xmm6, xmm5                   ; out lo
     paddw xmm6, LLO_ADD
@@ -778,15 +776,15 @@ loop2e:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -800,8 +798,8 @@ loop2e:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -820,7 +818,7 @@ loop2e:
     ; l[n] = src[2n] + ((h[n - 1] + h[n]) >> 1)
     movdqa xmm7, xmm5
     pslldq xmm7, 2
-    movd xmm6, ebx
+    movd xmm6, ebp
     por xmm7, xmm6
     paddw xmm5, xmm7
     psraw xmm5, 1
@@ -861,15 +859,15 @@ loop1e1:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -883,8 +881,8 @@ loop1e1:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -913,7 +911,7 @@ loop1e1:
     paddw xmm5, xmm1
 
     psrldq xmm2, 14
-    movd ebx, xmm2                      ; save hi
+    movd ebp, xmm2                      ; save hi
 
     movdqa [edx], xmm5                  ; out lo
 
@@ -930,15 +928,15 @@ loop2e1:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -952,8 +950,8 @@ loop2e1:
     movd xmm5, eax
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -973,14 +971,14 @@ loop2e1:
     ; l[n] = src[2n] + ((h[n - 1] + h[n]) >> 1)
     movdqa xmm7, xmm5
     pslldq xmm7, 2
-    movd xmm6, ebx
+    movd xmm6, ebp
     por xmm7, xmm6
     paddw xmm5, xmm7
     psraw xmm5, 1
     paddw xmm5, xmm1
 
     psrldq xmm2, 14
-    movd ebx, xmm2                      ; save hi
+    movd ebp, xmm2                      ; save hi
 
     movdqa [edx], xmm5                  ; out lo
 
@@ -998,15 +996,15 @@ loop2e1:
     movdqa xmm2, [esi + 16]
     movdqa xmm6, xmm1
     movdqa xmm7, xmm2
-    pand xmm1, [cdFFFF]
-    pand xmm2, [cdFFFF]
+    pand xmm1, [lsym(cdFFFF)]
+    pand xmm2, [lsym(cdFFFF)]
     packusdw xmm1, xmm2
     movdqa xmm2, xmm6                   ; src[2n + 1]
     movdqa xmm3, xmm7
     psrldq xmm2, 2
     psrldq xmm3, 2
-    pand xmm2, [cdFFFF]
-    pand xmm3, [cdFFFF]
+    pand xmm2, [lsym(cdFFFF)]
+    pand xmm3, [lsym(cdFFFF)]
     packusdw xmm2, xmm3
     movdqa xmm3, xmm6                   ; src[2n + 2]
     movdqa xmm4, xmm7
@@ -1020,8 +1018,8 @@ loop2e1:
     psrldq xmm5, 12
     pslldq xmm5, 12
     por xmm4, xmm5
-    pand xmm3, [cdFFFF]
-    pand xmm4, [cdFFFF]
+    pand xmm3, [lsym(cdFFFF)]
+    pand xmm4, [lsym(cdFFFF)]
     packusdw xmm3, xmm4
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -1040,7 +1038,7 @@ loop2e1:
     ; l[n] = src[2n] + ((h[n - 1] + h[n]) >> 1)
     movdqa xmm7, xmm5
     pslldq xmm7, 2
-    movd xmm6, ebx
+    movd xmm6, ebp
     por xmm7, xmm6
     paddw xmm5, xmm7
     psraw xmm5, 1
@@ -1080,9 +1078,9 @@ loop1f:
     punpcklbw xmm1, xmm0
     punpcklbw xmm2, xmm0
     punpcklbw xmm3, xmm0
-    psubw xmm1, [cw128]
-    psubw xmm2, [cw128]
-    psubw xmm3, [cw128]
+    psubw xmm1, [lsym(cw128)]
+    psubw xmm2, [lsym(cw128)]
+    psubw xmm3, [lsym(cw128)]
     psllw xmm1, 5
     psllw xmm2, 5
     psllw xmm3, 5
@@ -1114,8 +1112,8 @@ loop2f:
     movq xmm3, [esi + 64 * 1 * 2]       ; src[2n + 2]
     punpcklbw xmm2, xmm0
     punpcklbw xmm3, xmm0
-    psubw xmm2, [cw128]
-    psubw xmm3, [cw128]
+    psubw xmm2, [lsym(cw128)]
+    psubw xmm3, [lsym(cw128)]
     psllw xmm2, 5
     psllw xmm3, 5
     movdqa xmm4, xmm1
@@ -1147,7 +1145,7 @@ loop2f:
     movdqa xmm1, xmm3                   ; src[2n]
     movq xmm2, [esi + 64 * 1]           ; src[2n + 1]
     punpcklbw xmm2, xmm0
-    psubw xmm2, [cw128]
+    psubw xmm2, [lsym(cw128)]
     psllw xmm2, 5
     movdqa xmm4, xmm1
     movdqa xmm5, xmm2
@@ -1188,7 +1186,7 @@ set_quants_hi:
     movd xmm1, eax
     movdqa LHI_SFT, xmm1
     imul eax, 16
-    lea edx, [cwa0]
+    lea edx, [lsym(cwa0)]
     add edx, eax
     movdqa xmm1, [edx]
     movdqa LHI_ADD, xmm1
@@ -1199,7 +1197,7 @@ set_quants_lo:
     movd xmm1, eax
     movdqa LLO_SFT, xmm1
     imul eax, 16
-    lea edx, [cwa0]
+    lea edx, [lsym(cwa0)]
     add edx, eax
     movdqa xmm1, [edx]
     movdqa LLO_ADD, xmm1
@@ -1231,6 +1229,7 @@ PROC rfxcodec_encode_dwt_shift_x86_sse41
     movdqu [esp], xmm0
     ; save registers
     push ebx
+    RETRIEVE_RODATA
     push esi
     push edi
     push ebp
@@ -1385,5 +1384,4 @@ PROC rfxcodec_encode_dwt_shift_x86_sse41
     ; return value
     mov eax, 0
     ret
-    align 16
-
+END_OF_FILE